- Email: [email protected]
Adherence to and penetration of the intestinal epithelium by reovirus type 1 in neonatal mice
GASTROENTEROLOGY 1987:92:82-91
Adherence to and Penetration of the Intestinal Epithelium by Reovirus Type 1 in Neonatal Mice JACQUELINE L. WOLF, RINA DAMBRAUSKAS, ARLENE H. SHARPE, and JERRY S. TRIER Gastroenterology Division, Departments of Medicine and Pathology, Brigham and Women’s Hospital and Harvard Medical School; Department of Microbiology and Molecular Genetics, Harvard Medical School; and the Harvard Digestive Diseases Center, Boston, Massachusetts
In lo-day-old
suckling and adult mice, reovirus type 1 adheres selectively to and penetrates membranous epithelial (M) cells. To determine when M cells first appear, when they first transport reovirus, and if reovirus adheres to and is endocytosed by other epithelial cells in the first postnatal week, we examined neonatal mouse intestine by transmission electron microscopy after reovirus type 1 exposure. At 2 days M cells accounted for 0.9% of dome epithelial cells. By 9 days M cells had increased to 7.4%. Reovirus type 1 adherence to the surface of viJJus and dome epithelial cells showed marked variation in 2-6-day-old animals, but by 7 days only a few absorptive cell profiles had adherent reovirus. Adherence to >50% of M-cell profiles occurred in all but 2 animals, but adherence to the majority of Peyer’s patch absorptive cell profiles was present only in some 4- and S-day-old animals. Adherence to Q majority of undifferentiated cell profiles oc-
curred in some animals at all ages. Membranous epithelial cells endocytosed reovirus at all ages but only at 2 days did rare viJJus and dome absorptive cells endocytose reovirus into the apical cytoplasm.
Thus, adherence of reovirus to the apical surface of mucosal epithelial cells is nonselective in newborn mice but becomes more selective within the first postnatal
week with adherence
by day 7 to most
Received February 8, 1985. Accepted June 24, 1986. Address requests for reprints to: Jacqueline L. Wolf, M.D., Gastroenterology Division, Brigham and Women’s Hospital, 75 Francis Street, Boston, Massachusetts 02115. This study was supported in part by research grants 5 ROl AM36835 and 5 PO1 NS16998 from the National Institutes of Health. Dr. Wolf is the recipient of a Clinical Investigator Award [AM 01607) from the National Institutes of Health and a research grant from the National Foundation for Ileitis and Colitis. The authors thank Dr. Bernard N. Fields for helpful discussions and Kathryn K. McDade for technical assistance. 0 1987 by the American Gastroenterological Association 0016-5085/87/$3.50
M-cell profiles, to a substantial but variable number of undifferentiated cell profiles, but to few absorptive cell profiles. The initiation of systemic viral disease requires that viruses first penetrate an epithelial barrier of the host in order to gain access to the systemic circulation. A number of viruses, including the mammalian reoviruses, enter the systemic circulation by penetrating the intact gastrointestinal mucosal barrier (l-5). These viruses may replicate and produce disease locally and/or disseminate to distant organs via the lymphatics or blood supply. As viral penetration of the gut mucosa is an early and fundamental step in the pathogenesis of viral disease, we have been examining the route by which reoviruses cross the intestinal mucosa and the factors that influence this process. We have recently shown that in lo-day-old and adult mice, reovirus type I appears to penetrate the distal small intestinal mucosal barrier exclusively via the intestinal membranous epithelial (M) cell (3,4), a specialized epithelial cell that overlies Peyer’s patches (6-9).Adherence of reovirus type 1 to the M cell is selective; only rarely do type 1 reovirions adhere to absorptive cells at these ages. The M cell endocytoses adherent reovirions and transports them to the extracellular space and underlying lymphoid tissue of the patch (~,a). Reovirus type 1 is then disseminated to distant tissues (5) such as the brain where it produces ependymitis resulting in hydrocephalus [lo). Age influences the interaction of reovirus type 1 with tissues. After oral inoculation of mice, reovirus type I replicates more rapidly in the lung, liver, kidney, brain, and spleen, and reaches higher titers Abbreviation
used in this poper: M, membranous
epithelial.
REOVIRUS AND THE INTESTINE IN NEONATAL MICE
January 1987
in the heart, liver, kidney, and spleen in Zday-old than in lo-day-old mice (Sharpe AH and Fields BN, unpublished data). Similarly, when inoculated intracerebrally in mice, reovirus type 1 replicates more efficiently in the brains of l-day-old mice compared to mice aged 6 days or older (111. No difference was observed in the binding of reovirus type 1 to the surfaces of ependymal cells of newborn and adult animals (12).Whether reovirus type 1 penetrates the small intestine of younger neonates in the same manner as in lo-day-old and adult mice is unclear. Moreover, it is not known when M cells first appear, when they first transport reovirus, or if reovirus type 1 adheres to and is endocytosed by other intestinal epithelial cells, such as absorptive cells, which are known to endocytose macromolecules in the early neonatal period (13). To investigate these issues, we have studied the cellular composition of the epithelium overlying Peyer’s patches and the interaction of reovirus type 1 with the epithelial cells of the small intestine of mice 2-7 days of age.
Materials and Methods Animals Suckling BALBkByJ mice were obtained from Jackson Laboratories (Bar Harbor, Me.). Pups were suckled by natural or foster mothers until the time of study. During the study period each experimental mouse was kept in a separate cage covered with a filter cap. Virus Reovirus type 1 (strain Lang) was grown in L cells and purified according to previously described methods (14,15). The concentration of virus was determined by measuring the absorbance of the viral suspension at 260 nm (16). Purified virus was stored at 4°C and used within 2 mo of purification. Exposure
of Intestine
to Reovirus
BALBicByJ mice 2, 4, 5, 6, and 7 days old were anesthetized with ether and a laparotomy was performed. In 2-day-old mice, a single ligature was placed around the proximal colon and 0.05-0.12 ml of suspension containing reovirus type 1 (l-4 x 1Ol3 particles/ml) or phosphatebuffered saline was instilled proximally into the small intestines. In &7-day-old mice, closed loops were made with the distal ligature placed around the proximal colon or distal ileum and the proximal ligature placed in the ileum in such a way as to include one or more Peyer’s patches within the loop. Care was taken not to injure the mesenteric blood supply. Loops were filled with 0.02-0.07 ml of reovirus suspension or phosphate-buffered saline. The jejunum and ileum of 2-day-old mice and the closed ileal loops from the older mice were removed 0.5 or 1 h after inoculation. Because Peyer’s patches were difficult to
83
identify in 2-day-old mice, both jejunum and ileum were used, In older animals, patches could be readily identified, thus only ileal loops were inoculated and examined. For each age, 4-15 experimental mice were exposed to reovirus. At each age an additional 1-5 mice were exposed to virus-free buffer and their intestines were examined to assure all litters were free of enteric virus infection. Peyer’s patches were removed from the loops, pinned mucosal side up to paraffin in a petri dish, and fixed for transmission electron microscopy. If no Peyer’s patches could be identified, as was the case in some 2-day-old mice, random segments of bowel were removed and put directly into fixative. Microscopy Samples for transmission electron microscopy were fixed in Karnovsky’s solution (17) for 2 h, postfixed in 1% aqueous 0~0, for 1 h, dehydrated, and embedded in Epon 812 using previously described methods (4). Thin sections of Peyer’s patches identified in l-2-pm toluidine blue-stained sections of intestinal tissue were cut on a Sorvall MTZ-B ultramicrotome, stained with uranyl acetate and lead citrate, and examined with a Philips 300 electron microscope (Philips Electronic Instruments, Mahwah, N.J.). The percentage of each epithelial cell type in epithelium overlying Peyer’s patches and the percentage of each cell type in epithelium overlying villi and Peyer’s patches with adherent reovirus were quantitated for each animal studied. The mean and standard error of the mean for animals of a given age were then determined. To prevent scoring of more than one sectioned profile of each cell, only a single section from each grid was scored except for areas obscured by a bar of the copper grid. These obscured areas were counted on an adjacent section. When more than one section was scored from a given tissue block, only sections obtained at depth intervals of -15 pm were used. Cells with two or more reovirions closely applied to the microvillus surface were scored as positive for reovirus adherence. Statistics The results are expressed as mean * standard error. Multiple sample comparisons of Kruskal-Wallis and Newman-Keul were done to compare age-related changes of cell type and adherence of reovirus. The variances were compared by Levy’s multiple comparisons.
Results Peyer’s patches were identified on the serosal surface of the antimesenteric side of the small intestine as a bulge or an area having a network of anastomosing vessels associated with lymphoid follicles. Peyer’s patches were easily seen with a dissecting microscope in mice 4 days of age and older but were only occasionally recognized in this manner in mice 2 days of age. In most 2-day-old mice, light microscopic examination of serial l-2-pm sec-
84
WOLF ET AL
tions was required to identify patches. From the 29 blocks of tissue that were studied from fifteen &dayold mice, 10 Peyer’s patches from 7 different mice were identified. The Peyer’s patches observed in 2-day-old mice, previously termed “anlage” patches (18), were small and appeared immature (Figure la). The most abundant cell types were large and medium-sized lymphocytes, although some macrophages were present. Very few small lymphocytes were seen. Between postnatal days 4 and 7, the size of Peyer’s patches and number of small lymphocytes in most Peyer’s patches gradually increased (Figure lb). All 10 immature Peyer’s patches identified in the &day-old mice were examined by transmission electron microscopy. Nine of 10 patches were from mice inoculated with reovirus. The majority (58%) of cells overlying domes of these immature patches were undifferentiated epithelial cells (Table 1). These cells had relatively short microvilli, a poorly developed terminal web, only sparse elements of the apical tubulovesicular system, many free ribosomes, and few elements of formed endoplasmic reticulum (Figure 2). In contrast, characteristic absorptive cells of these neonatal mice, which constituted 38.5% of the dome epithelium (Table l), had longer microvilli, a well-developed apical tubulovesicular system, fewer free ribosomes, and more formed endoplasmic reticulum (Figure 31. Tuft and goblet cells, although uncommon, were seen occasionally. A total of six M cells was identified in the epithelium overlying three of the 10 Peyer’s patches obtained from two 2-day-old mice. These M cells had structural features similar to those seen in older animals; both mature and immature forms were present (6-9,19). The M cells had short, irregular, branched and wide microvilli and a poorly developed terminal web. However, only one M-cell profile had the typical central hollow that contained mononuclear cells (6-9,19). In other M-cell profiles, mononuclear cells were adjacent to the basal cytoplasm, which they often indented (Figure 4). As the mice aged, the mean percentage of undifferentiated cells overlying Peyer’s patch domes decreased from 58% at 2 days of age to 41% at 7 days of age, whereas the percentage of absorptive cells increased from 38.5% at 2 days of age to 48% at 7 days of age (Table 1). The mean percentage of M cells increased from 0.9% of the cells overlying Peyer’s patch domes at 2 days to 7.4% at 7 days. The percentage of goblet and tuft cells remained essentially unchanged between 2 and 7 days after birth. However, in mice 2-6 days of age, there was a large variation from patch to patch in the percentage of undifferentiated cells in the epithelium overlying sectioned patch profiles. This variation occurred
GASTROENTEROLOGY
Figure
1
Vol. 92, No. 1
Low-magnification electron micrographs of lymphoid follicles of Peyer’s patches from a &day-old (a) and a 6-day-old (b) BALBicByJ suckling mouse. The dark lines represent the bars of the copper grid. The greater size and increased cellularity at 6 days is evident (X600).
REOVIRUS AND THE INTESTINE
January 1987
Table
1. Cellular
Composition
of the Epithelium
IN NEONATAL
85
MICE
of Peyer’s Patches Mouse age [days]
Undifferentiated cells” Absorptive cells” M cells” Goblet cells” Tuft cells” Total cells
2
4
57.8 ?z 8.3 38.5 -r-9.4 0.9 k 0.9 1.9 2 1.0 1.2 k 0.9 197
57.8 2 7.8 37.8 2 7.3 3.0 +- 1.5 1.5 2 0.8 0.2 2 0.2 301
6
5 48.8 2 39.0 ? 4.8 + 5.2 f 2.6 2 208
7.3 5.4 2.3 1.1 1.5
38.5 ? 51.5 k 3.6 k 5.9 + 0.5 + 254
7 7.8 8.0 1.6 1.7 0.5
41.2 k 47.8 t 7.4 t 2.1 t 1.7 t 205
1.2 2.8 2.7 1.1 0.6
a Mean percentage 2 SEM. The counted cells were from the following number of patches and (mice]: day 2,5 (4); day 4, 6 (4): day 5.4 (4); day 6, 6 (4);day 7, 6 (5).
even among patches from the same animal. This is not surprising, as the intestine does not mature at the same rate in all mice nor do all patches mature in synchrony in a given mouse. By 7 days, the variation in the percentage of undifferentiated cells overlying patches from different animals decreased from a mean SE of 7.8 in the 2-6-day-old animals to 1.2 in the 7-day-old animals. This difference in variance was statistically significant (p < 0.01). There was also substantial variation in the number and percentage of M cells in sectioned profiles of patches in young neonates. At 2 days of age, only one M cell (3.6% of dome cells on that patch) was detected in the five scored patches from 4 animals. In &day-old mice, the number of detectable M cells in scored patch profiles varied from 7.1% to 0%. At 7 days, >lO% of the cell profiles from patches of 2 mice were M cells, whereas in another mouse 22% and 2.5% of dome epithelial cell profiles in each of two patches were M cells. In 2 other ‘I-day-old mice, 0% and 1.9% of the dome epithelial cell profiles were M cells. As a result of this variation, differences in the percentage of undifferentiated, absorptive, and M cells in patch profiles of mice 2-7 days of age did not achieve significance when subjected to rigorous statistical analysis. Reovirus adhered to the apical surface of >50% of M-cell profiles in all mice except for 2, in which adherence to 0 of 1 M-cell profile in 1 mouse and to 40% of the M-cell profiles in the other was demonstrated. Of the six M-cell profiles visualized in 2-dayold patches, five had adherent reovirus on their apical surface. Reovirus adherence to undifferentiated cells overlying Peyer’s patches showed large variations from patch to patch and animal to animal. At 2 and 4 days, a mean of 53% k 13% and 56% +- 8% of undifferentiated cell profiles had adherent reovirus, whereas from 5 to 7 days the mean percentage with adherent reovirus was 31%-36.5% (Figure 5). These differences were not statistically significant. Reovirus adherence to absorptive cell profiles overlying Peyer’s patch domes varied from a mean of
15% at 2 days to 34% at 4 days, and then decreased to 11% at 7 days. Variation of adherence among
individual mice was large, resulting in large standard errors and considerable overlap between
Figure
2. Undifferentiated cell with adherent reovirus (arrows] on the apical surface located in the epithelium overlying the dome of a Peyer’s patch of a Z-day-old BALBicByJ mouse 30 min after intrajejunal inoculation with reovirus type 1. Note the numerous free ribosomes and the absence of elements of an apical tubulovesicular system (X 18,500).
86
WOLF ET AL.
Figure
3
Ileal epithelium from a Z-day-old BALB/cByJ mouse 1 h after intrajejunal inoculation with reovirus type 1. a. A villus absorptive cell with reovirus adherent to the microvilli (short arrows) and within apical vesicles (long arrows) (X10,800). b. Peyer’s patch absorptive cell cytoplasm with reovirus (arrow] within a lysosomelike body (X72,900).
GASTROENTEROLOGY
Vol. 92, No. 1
groups (Figure 5). In contrast to the adherence of reovirus to the majority of M-cell profiles, in only three patches from two 4-day-old animals and a single patch from one 5-day-old animal did >50% of absorptive cell profiles have adherent reovirus. Adherence of reovirus to the apical surfaces of the villus absorptive cell profiles showed large variations among young neonates with means of 36% + 20% at 2 days and 54% k 11% at 4 days of age. The number of villus absorptive cell profiles with adherent reovirus decreased dramatically by 7 days; only 3% + 2.8% had adherent reovirus. Because of the overall large variance, however, the decrease in adherence with age did not achieve significance when subjected to rigorous statistical analysis (Figure 5). The decrease in the variance of adherence at 7 days was significant (p < 0.01). Adherence of reovirus to the apical cell surface did not correlate with endocytosis of reovirions into the epithelial cytoplasm in 2-day-old mice. Evidence of endocytosis by undifferentiated cells of reovirus type 1 was not seen although virus adhered to sectioned profiles of the majority of these cells. In contrast, in 12 of the many hundreds of absorptive cell profiles surveyed, reovirus type 1 was found within vesicles and/or lysosomelike bodies in the apical cytoplasm. Ten of these overlaid villi and two overlaid Peyer’s patches (Figures 3a and 3b). No differences were detected in the structural features of those few absorptive cells that endocytosed type 1 reovirions and the many that did not. At 2 days of age, multiple sections of each of six M cells showed that two had endocytosed reovirus into the apical cytoplasm (Figure 4).Within the M-cell cytoplasm, reovirus was found in apical vesicles and in a multivesicular body. However, at 2 and 4 days of age 0.5and 1 h after inoculation, no reovirions were seen in the basal two-thirds of cytoplasm of any epithelial cell, in intercellular spaces, or in the underlying lamina propria. In mice older than 2 days, reovirus was found within the cytoplasm only of M cells except for a single absorptive cell profile from a &day-old mouse in which one virion was found in an apical vesicle close to the microvillus surface. It could not be determined whether this was an endocytic vesicle or whether it represented a tangentially sectioned pit that communicated with the intestinal lumen out of the plane of section. In mice 5 days of age and older, evidence that M cells both endocytosed and transported reovirus type 1 across the epithelium was observed. Reovirions were found on the M-cell apical surface, within M-cell cytoplasmic vesicles, and in the intercellular space underlying the M cell (Figure 6). No reoviruslike structures were seen in any of the
January
1987
REOVIRUS
AND THE INTESTINE
IN NEONATAL
MICE
87
FYigure 4. Ileal epithelium overlying the dome of an immature Peyer’s patch from a Z-day old BALB/cByJ mouse obtained 1 h after intrajejunal inoculation with reovirus type 1. Reovirus is adherent to the apical surfaces of a tuft (T) and M (M) cell, but not to the apical surface of an adjacent absorptive (A) cell. and is found within apical cytoplasmic vesicles of the M cell (arrows) (x11.800).
88
WOLF
ET AL.
GASTROENTEROLOGY
Vol. 92. No. 1
70
In .-2 > $ a
60 f 50
.
t
z ? g 40 a c .t 3 30 UY = S z
20
8
130 169 102 95
N= AGE (days I=
2
4
Undifferentiated
5
6
86
59
7
2
Cells on
4
5
Absorptive
Peyer ‘s Patches Figure
121 83
Peyer’s
133
97
6
7
Cells
on
Patches
277
-
293 285 35 I 240
24567 Absorptive
Cells on
Villi
5. Adherence of reovirus type 1 to undifferentiated cells and absorptive cells overlying the domes of Peyer’s patches absorptive cells overlying villi from 2-7-day-old BALBicByJ mice killed 0.5-l h after inoculation with reovirus type standard error of the mean is indicated. The total number (N) of each cell type counted at a given age is shown.
control mice that had phosphate-buffered stilled into the intestinal lumen.
saline
in-
Discussion We have previously documented that reovirus type 1 adheres selectively to the apical surface of the M cell and is transported through its cytoplasm in endocytic vesicles to the underlying Peyer’s patch lymphoid follicle from which it is then disseminated to distant organs in mice 10 days of age and older (3-5). However, it was not known whether this occurred in younger neonatal mice, the time at which reovirus type 1 grows most avidly in many tissues including the brain (11; Sharpe AH and Fields BN, unpublished data). As it also was not known when M cells first appear and first transport reovirus during intestinal morphogenesis in mice, we determined in suckling mice 2-7 days of age (a) whether M cells are present, (b) the pattern of reovirus type 1 adherence to intestinal epithelial cells, and (c) which epithelial cell types endocytose reovirus type 1.
and to 1. One
Small intestinal M cells are not present on villi but are confined to the epithelium overlying the dome of lymphoid follicles (6-9). In previous studies, rudimentary Peyer’s patches were observed in the small intestine of rats 2 days before birth (20,21) and at birth in several strains of mice including BALB/cJ mice (18,22-25). These patches in newborn mice are not fully developed; they contain endothelial, histiocytic, reticular, and large lymphoid cells but few small lymphocytes at birth. However, over the next 3-4 days, they become populated with large numbers of small lymphocytes (18,22-25). Compartmentalization of Peyer’s patches in suckling rats into well-defined follicular and interfollicular zones does not occur until 12-14 days after birth (26). Membranous epithelial cells have been noted overlying the bursa of Fabricius in 15day-old chicken embryos and at birth in rabbits (7). A single putative M cell has been illustrated overlying a submucosal lymphoid cell collection in the small intestine of a 17-wk-old human fetus (27), and M cells were reported to be present but were not illustrated overly-
lanuary 1987
Figure
REOVIRUS
AND THE INTESTINE
IN NEONATAL
6. Epithelium overlying a Peyer’s patch from a &day-old BALBkByJ mouse 1 h after inoculation with reovirus are seen adherent to the luminal surface of an M cell, within vesicles of the M cell, and within the extracellular (X 11,600). Inset: higher magnification of one of the extracellular virions (x93.400).
ing lymphoid follicles in the small intestine of newborn rats (20). Our findings confirm that aggregates of lymphoid cells that probably represent immature Peyer’s patches are present in the small intestine in the early neonatal period in mice. We also found that a small fraction (Cl%) of the epithelial cells that overlie these immature Peyer’s patches 2 days after birth are differentiated M cells but over the next 5 days of development, the mean frequency of M cells in dome epithelium increases over eightfold. This increase with age of the percentage of M cells on the apical surface of the domes of Peyer’s patches does not appear to occur in all animals of a given age at the same time. We noted considerable variation in Mcell frequency over patches among animals of the same age and even among individual patches from the same animal. Undifferentiated cells are the most abundant cell type overlying domes of immature patches in z-day-
MICE
89
type 1. Virions space (arrows)
old mice. However, as the patches mature over the next 5 days, the percentage of undifferentiated cells appears to decrease progressively as the percentage of absorptive cells increases. However, there is considerable variation among mice of the same age and among different patches from the same mouse, suggesting that maturation of the epithelium overlying patches does not proceed at the same rate and probably reflects the degree of maturity of each individual patch. Reovirus type 1 adheres selectively to M cells and a few tuft cells in lo-day-old suckling mice but not to absorptive or goblet cells overlying ileal Peyer’s patch domes (3,~). In adult mice, however, in addition to adherence of virions to most M cells and a few tuft cells, adherence of virions to a variable but small percentage of patch absorptive and goblet cells was observed (4). No adherence to villus epithelium was observed in either lo-day-old suckling or adult mice. In the present study, we found that reovirus
90 WOLFETAL.
type 1 adhered consistently to >50% of sectioned M-cell profiles in mice from 2 to 7 days old in all but two patches, one of which had only a single M cell. On the other hand, there was also adherence to a large percentage of sectioned profiles of undifferentiated cells overlying Peyer’s patches in young neonates. At 2 days, reovirus adhered to a small percentage of Peyer’s patch absorptive cells. At 4 days, adherence to absorptive cells was maximal both over patches and on villi. After 4 days, there was a tendency for the adherence of type 1 reovirus to absorptive cells to decrease with advancing age; however, this did not achieve statistical significance because there was substantial variation among animals of a given age. Again, differences in the rate of intestinal maturation from animal to animal and the presence of Peyer’s patches at various stages of development in animals of a given age may be responsible for these large variances. However, by 7 days, although virus adhered to the majority of sectioned M-cell profiles, mean adherence to sectioned absorptive cell profiles overlying patches and villi was only 10% and 3% respectively, strongly suggesting the emergence of increasing selectivity of type 1 reovirus adherence similar to that previously documented by us in lo-day-old suckling and adult mice (3,4,19). Several factors may influence the emergence of selective adherence of type 1 reovirus to M cells on dome epithelium in mice during the first 10 neonatal days. Undifferentiated cells, which bind reovirus type 1 in neonates and which retain some capacity to bind reovirus type 1 even in adult mice (Bye WA and Trier JS, unpublished data), ultimately disappear from the overlying dome epithelium as Peyer’s patches mature. Instead, they become segregated to the developing crypts that eventually encircle the domes of Peyer’s patches much as crypts encircle villi (19,28). That most patch and villus absorptive cells fail to bind reovirus type 1 by 7-10 days after birth may reflect alterations in the composition of the glycoconjugates of the apical glycocalyx. Changes in surface glycoconjugates may also account for the apparent difference in adherence to patch absorptive cells in 2- versus 4-day-old mice. It is well established that the glycocalyx of the brush border of intestinal absorptive cells in adult animals changes dramatically during normal maturation from precursor undifferentiated crypt cells (29). In suckling rats, the binding of lectins to glycoconjugates on the apical surface of villus absorptive cells changes dramatically during the first 2 wk of postnatal development, suggesting a change in the exposed sugar and sialic acid residues in the glycocalyx (30).Alterations in the incorporation of precursor sugars into glycoconjugates of the microvillus membrane of rat intestinal
GASTROENTEROLOGYVol. 92,No.l
epithelial cells have also been noted during postnatal development (31,32). Thus, although the glycoconjugates in the apical membrane of epithelial cells overlying the domes of Peyer’s patches in mice have not yet been directly examined during postnatal development, there is abundant evidence that epithelial cell surface glycoconjugates elsewhere in the small intestine of the closely related rodent, the rat, change dramatically with maturation during the early postnatal period. Factors other than glycocalyx composition may be important in determining reovirus adherence during intestinal development. Changes in the intraluminal environment during postnatal maturation of the stomach and small intestine may also influence the evolution of reovirus type 1 adherence to intestinal epithelial cells. Treatment of cultured L cells with pronase decreases binding of type 1 reovirus by -50% (33). Intraluminal trypsin activity is very low in the intestine of newborn rats but increases on day 4 (34). The gastric pH in rats decreases considerably as acid secretion increases during the first 10 postnatal days (35). Changes in protease activity and intraluminal pH may result in structural alterations of the brush border glycocalyx of intestinal epithelial cells or reovirus surface proteins altering viral adherence and infectivity (15,33,36,37). Endocytosis of reovirus type 1 by absorptive cells was observed with certainty only in 2-day-old mice, in which it was a rare phenomenon. However, whether endocytosis of reovirus type 1 into absorptive cells in 2-day-old mice reflects internalization of reovirions adherent to the absorptive cell apical membrane or nonspecific fluid phase endocytosis, which characterizes macromolecular uptake in mouse ileal absorptive cells from birth until weaning (131, was not established by our studies. Endocytosed virions were confined to vesicles and lysosomelike structures in the apical cytoplasm and were never observed in the basal two-thirds of absorptive cells, in intercellular spaces between absorptive cells, or in the underlying lamina propria. Whether these internalized reovirions would ultimately be degraded, replicated, or transported across the cell is unknown. However, from our morphologic studies, it seems unlikely that absorptive cells represent an important site for reovirus type 1 penetration of the ileal mucosa, especially in mice 24 days old. Although uptake of virions by absorptive cells more than 1 h after inoculation cannot be ruled out in these neonatal mice, our previous studies in mice ~10 days old (4)showed that reovirus type 1 was not endocytosed by absorptive cells within 4 h after inoculation. In contrast, M cells endocytosed reovirus type 1 at all ages examined and appear to be the major if not the only site for penetration of the intestinal mucosal barrier by this virus.
January
REOVIRUS
1987
The presence of M cells in Z-day-old mice suggests that very early in the postnatal period reovirus can be transported across the epithelial barrier of the intestine to the underlying lymphoid tissue. However, because the patch is immature, the ensuing immunologic response may differ in neonatal compared with older suckling and weaned mice and may permit the more exuberant growth of reovirus observed in tissues distant from the intestine during the early neonatal period (Sharpe AH and Fields BN, unpublished data). Hence, it appears likely that the differences in growth of reovirus type 1 observed in tissues distant from the intestine as the mouse ages reflect events occurring at a step beyond actual penetration of the epithelial barrier.
of virions. 17.
18.
19.
20.
21.
22.
References concept of poliomyelitis infection. Sci1. Bodian D. Emerging ence 1955;122:105-8. BR. Mims CA, Sambrook J, White DO. 2. Fenner F. McAuslan 1974: The biology of animal viruses. New York: Academic, 356-9. R, et al. Intestinal M cells: a 3. Wolf JL, Rubin DH, Finberg pathway for entry of reovirus into the host. Science 1981; 212:471-2. RS. Finberg R, Dambrauskas R, Fields BN, 4. Wolf JL, Kauffman Trier JS. Determinants of reovirus interaction with the intestinal M cells and absorptive cells of murine intestine. Gastroenterology 1983;85:291-300. RS, Wolf JL, Finberg R, Trier JS, Fields BN. The e1 5. Kauffman protein determines the extent of spread of reovirus from the gastrointestinal tract of mice. Virology 1983;124:403-10. 6. Wolf JL, Bye WA. The membranous epithelial (M) cell and the mucosal immune system. Ann Rev Med 1984;35:95-112. DE, Cooper MD. Pinocytosis by epithelium associ7. Bockman ated with lymphoid follicles in the bursa of Fabricius, appendix, and Peyer’s patches. An electron microscopic study. Am J Anat 1973;136:455-77. cell specialization within 8. Owen RL, Jones AL. Epithelial human Peyer’s patches: an ultrastructural study of intestinal lymphoid follicles. Gastroenterology 1974;66:189-203. 9. Owen RL. Sequential uptake of horseradish peroxidase by lymphoid follicle epithelium of Peyer’s patches in the normal unobstructed mouse intestine: an ultrastructural study. Gastroenterology 1977;72:440-51. 10. Margolis G, Kilham L. Hydrocephalus in hamsters, ferrets, rats and mice following inoculations with reovirus type 1: pathologic studies. Lab Invest 1969;21:189-98. 11. Tardieu M, Powers ML, Weiner HL. Age dependent susceptibility to reovirus type 3 encephalitis: role of viral and host factors. Ann Neural 1983;13:602-7. 12. Tardieu M, Weiner HL. Viral receptors on isolated murine and human ependymal cells. Science 1982:215:419-22. 13. Clark SL Jr. The ingestion of proteins and colloidal materials by columnar absorptive cells of the small intestine in suckling rats and mice. J Biophys Biochem 14. Sharpe AH, Fields BN. Reovirus
Cytol 1959;5:41-50. inhibition of cellular
RNA
and protein synthesis: role of the S, gene. Virology 1982; 122:381-91. 15. Drayna D, Fields BN. Genetic studies on the mechanism of chemical and physical inactivation of reovirus. J Gen Virol 1982;83:149-59. 16. Smith
RE, Zweerink
HJ, Joklik
WK. Polypeptide
components
AND THE INTESTINE
top
component
and
IN NEONATAL
cores
MICE
of reovirus
type
91
3.
Virology 1969;39:791-810. Karnovsky MJ. A formaldehyde-glutaraldehyde fixative of high osmolality for use in electron microscopy. J Cell Biol 1965;27:137A-8A3. Miller JJ, Studies of the phylogeny and ontogeny of the specialized lymphatic tissue venules. Lab Invest 1969:Zl: 484-90. Bye WA. Allan CH, Trier JS. Structure, distribution, and origin of M cells in Peyer’s patches of mouse ileum. Gastroenterology 1984;86:789-801. Wilders MM, Sminia T, Janse EM. Ontogeny of non-lymphoid and lymphoid cells in the rat gut with special reference to large mononuclear Ia-positive dendritic cells. Immunology 1983:50:303-31. Hummel KP. The structure and development of the lymphatic tissue in the intestine of the albino rat. Am J Anat 1935; 57:351-83. Waksman BH. The homing pattern of thymus-derived lymphocytes in calf and neonatal mouse Peyer’s patches. J Im-
munol 1973;111:878-84. SH. Weissman IL. Lymphoid tissue architecture: II. 23. Friedberg Ontogeny of peripheral T and B cells in mice: evidence against Peyer’s patches as the site of generation of B cells. J Immunol 1974;113:1477-92. 24. Joel DD. Hess MW. Cottier H. Thymic origin of lymphocytes in developing Peyer’s patches of newborn mice. Nature New Biol 1971:231:24-5. and pattern of 25. Joel DD, Hess MW. Cottier H. Magnitude thymic lymphocyte migration in neonatal mice. J Exp Med 1972;135:907-23. of Peyer’s patches 26. Sminia T, Janse EM, Plesch BEC. Ontogeny of the rat. Anat Ret 1983:207:309-16. cell types in the human fetal 27. Moxey PC, Trier JS. Specialized small intestine. Anat Ret 1978;191:269-86. in lymphoid 28. Bhalla DK, Owen RL. Cell renewal and migration follicles of Peyer’s patches and cecum. An autoradiographic study in mice. Gastroenterology 1982:82:232-42. morphology of the mucosa of 29. Trier JS. Madara JL. Functional the small intestine. In: Johnson LR. Christensen J, Grossman MI, Jacobson ED, Schultz SG. eds. Physiology of the gastrointestinal tract. New York: Raven, 1981:925-61. ML. Cell surface components of intes30. Etzler ME, Branstrator tinal epithelial cells and their relationship to cellular differentiation In: Development of mammalian absorptive processes. Ciba Foundation Symposium 70 (new series). New York: Excerpta Medica, 1979:51-68. A, Walker WA. Differences in microvil31. Bresson JL, Hercovics lus membrane glycoprotein fucosylation with age and differentiation (abstr). Gastroenterology 1982;82:1025. and transport of glycoproteins 32. Yen K-Y, Moog F. Biosynthesis in the small intestinal epithelium of rats: I. Developmental change and effect of hypophysectomy. Dev Biol 1984;lOl: 446-62. of 33. Epstein RL, Powers ML, Rogart RB, Weiner HL. Binding ‘*‘I-labeled reovirus to cell surface receptors. Virology 1984;133:46-55. P, Deschodt-Lanckman M. Camus J, Bruylands J. 34. Robberecht Christophe J. Pancreatic hydrolases from birth to weaning and dietary adaptation after weaning. Am J Physiol 1971; 221:376-81. HF. Ultrastructure and function of gastric parietal 35. Helander cells in the rat during development. Gastroenterology 1969;56:35-52. RS, Schaffer FL. Enzymatic enhancement of infec36. Spendlove tivity of reovirus. J Bacterial 1965;89:597-602. 37. Hauri H-P, Quaroni A, Isselbacher KJ. Biogenesis of intestinal plasma membrane: post-translational route and cleavage of sucrase-isomaltase. Proc Nat1 Acad Sci USA 1979;76:5183-6.
Adherence to and Penetration of the Intestinal Epithelium by Reovirus Type 1 in Neonatal Mice JACQUELINE L. WOLF, RINA DAMBRAUSKAS, ARLENE H. SHARPE, and JERRY S. TRIER Gastroenterology Division, Departments of Medicine and Pathology, Brigham and Women’s Hospital and Harvard Medical School; Department of Microbiology and Molecular Genetics, Harvard Medical School; and the Harvard Digestive Diseases Center, Boston, Massachusetts
In lo-day-old
suckling and adult mice, reovirus type 1 adheres selectively to and penetrates membranous epithelial (M) cells. To determine when M cells first appear, when they first transport reovirus, and if reovirus adheres to and is endocytosed by other epithelial cells in the first postnatal week, we examined neonatal mouse intestine by transmission electron microscopy after reovirus type 1 exposure. At 2 days M cells accounted for 0.9% of dome epithelial cells. By 9 days M cells had increased to 7.4%. Reovirus type 1 adherence to the surface of viJJus and dome epithelial cells showed marked variation in 2-6-day-old animals, but by 7 days only a few absorptive cell profiles had adherent reovirus. Adherence to >50% of M-cell profiles occurred in all but 2 animals, but adherence to the majority of Peyer’s patch absorptive cell profiles was present only in some 4- and S-day-old animals. Adherence to Q majority of undifferentiated cell profiles oc-
curred in some animals at all ages. Membranous epithelial cells endocytosed reovirus at all ages but only at 2 days did rare viJJus and dome absorptive cells endocytose reovirus into the apical cytoplasm.
Thus, adherence of reovirus to the apical surface of mucosal epithelial cells is nonselective in newborn mice but becomes more selective within the first postnatal
week with adherence
by day 7 to most
Received February 8, 1985. Accepted June 24, 1986. Address requests for reprints to: Jacqueline L. Wolf, M.D., Gastroenterology Division, Brigham and Women’s Hospital, 75 Francis Street, Boston, Massachusetts 02115. This study was supported in part by research grants 5 ROl AM36835 and 5 PO1 NS16998 from the National Institutes of Health. Dr. Wolf is the recipient of a Clinical Investigator Award [AM 01607) from the National Institutes of Health and a research grant from the National Foundation for Ileitis and Colitis. The authors thank Dr. Bernard N. Fields for helpful discussions and Kathryn K. McDade for technical assistance. 0 1987 by the American Gastroenterological Association 0016-5085/87/$3.50
M-cell profiles, to a substantial but variable number of undifferentiated cell profiles, but to few absorptive cell profiles. The initiation of systemic viral disease requires that viruses first penetrate an epithelial barrier of the host in order to gain access to the systemic circulation. A number of viruses, including the mammalian reoviruses, enter the systemic circulation by penetrating the intact gastrointestinal mucosal barrier (l-5). These viruses may replicate and produce disease locally and/or disseminate to distant organs via the lymphatics or blood supply. As viral penetration of the gut mucosa is an early and fundamental step in the pathogenesis of viral disease, we have been examining the route by which reoviruses cross the intestinal mucosa and the factors that influence this process. We have recently shown that in lo-day-old and adult mice, reovirus type I appears to penetrate the distal small intestinal mucosal barrier exclusively via the intestinal membranous epithelial (M) cell (3,4), a specialized epithelial cell that overlies Peyer’s patches (6-9).Adherence of reovirus type 1 to the M cell is selective; only rarely do type 1 reovirions adhere to absorptive cells at these ages. The M cell endocytoses adherent reovirions and transports them to the extracellular space and underlying lymphoid tissue of the patch (~,a). Reovirus type 1 is then disseminated to distant tissues (5) such as the brain where it produces ependymitis resulting in hydrocephalus [lo). Age influences the interaction of reovirus type 1 with tissues. After oral inoculation of mice, reovirus type I replicates more rapidly in the lung, liver, kidney, brain, and spleen, and reaches higher titers Abbreviation
used in this poper: M, membranous
epithelial.
REOVIRUS AND THE INTESTINE IN NEONATAL MICE
January 1987
in the heart, liver, kidney, and spleen in Zday-old than in lo-day-old mice (Sharpe AH and Fields BN, unpublished data). Similarly, when inoculated intracerebrally in mice, reovirus type 1 replicates more efficiently in the brains of l-day-old mice compared to mice aged 6 days or older (111. No difference was observed in the binding of reovirus type 1 to the surfaces of ependymal cells of newborn and adult animals (12).Whether reovirus type 1 penetrates the small intestine of younger neonates in the same manner as in lo-day-old and adult mice is unclear. Moreover, it is not known when M cells first appear, when they first transport reovirus, or if reovirus type 1 adheres to and is endocytosed by other intestinal epithelial cells, such as absorptive cells, which are known to endocytose macromolecules in the early neonatal period (13). To investigate these issues, we have studied the cellular composition of the epithelium overlying Peyer’s patches and the interaction of reovirus type 1 with the epithelial cells of the small intestine of mice 2-7 days of age.
Materials and Methods Animals Suckling BALBkByJ mice were obtained from Jackson Laboratories (Bar Harbor, Me.). Pups were suckled by natural or foster mothers until the time of study. During the study period each experimental mouse was kept in a separate cage covered with a filter cap. Virus Reovirus type 1 (strain Lang) was grown in L cells and purified according to previously described methods (14,15). The concentration of virus was determined by measuring the absorbance of the viral suspension at 260 nm (16). Purified virus was stored at 4°C and used within 2 mo of purification. Exposure
of Intestine
to Reovirus
BALBicByJ mice 2, 4, 5, 6, and 7 days old were anesthetized with ether and a laparotomy was performed. In 2-day-old mice, a single ligature was placed around the proximal colon and 0.05-0.12 ml of suspension containing reovirus type 1 (l-4 x 1Ol3 particles/ml) or phosphatebuffered saline was instilled proximally into the small intestines. In &7-day-old mice, closed loops were made with the distal ligature placed around the proximal colon or distal ileum and the proximal ligature placed in the ileum in such a way as to include one or more Peyer’s patches within the loop. Care was taken not to injure the mesenteric blood supply. Loops were filled with 0.02-0.07 ml of reovirus suspension or phosphate-buffered saline. The jejunum and ileum of 2-day-old mice and the closed ileal loops from the older mice were removed 0.5 or 1 h after inoculation. Because Peyer’s patches were difficult to
83
identify in 2-day-old mice, both jejunum and ileum were used, In older animals, patches could be readily identified, thus only ileal loops were inoculated and examined. For each age, 4-15 experimental mice were exposed to reovirus. At each age an additional 1-5 mice were exposed to virus-free buffer and their intestines were examined to assure all litters were free of enteric virus infection. Peyer’s patches were removed from the loops, pinned mucosal side up to paraffin in a petri dish, and fixed for transmission electron microscopy. If no Peyer’s patches could be identified, as was the case in some 2-day-old mice, random segments of bowel were removed and put directly into fixative. Microscopy Samples for transmission electron microscopy were fixed in Karnovsky’s solution (17) for 2 h, postfixed in 1% aqueous 0~0, for 1 h, dehydrated, and embedded in Epon 812 using previously described methods (4). Thin sections of Peyer’s patches identified in l-2-pm toluidine blue-stained sections of intestinal tissue were cut on a Sorvall MTZ-B ultramicrotome, stained with uranyl acetate and lead citrate, and examined with a Philips 300 electron microscope (Philips Electronic Instruments, Mahwah, N.J.). The percentage of each epithelial cell type in epithelium overlying Peyer’s patches and the percentage of each cell type in epithelium overlying villi and Peyer’s patches with adherent reovirus were quantitated for each animal studied. The mean and standard error of the mean for animals of a given age were then determined. To prevent scoring of more than one sectioned profile of each cell, only a single section from each grid was scored except for areas obscured by a bar of the copper grid. These obscured areas were counted on an adjacent section. When more than one section was scored from a given tissue block, only sections obtained at depth intervals of -15 pm were used. Cells with two or more reovirions closely applied to the microvillus surface were scored as positive for reovirus adherence. Statistics The results are expressed as mean * standard error. Multiple sample comparisons of Kruskal-Wallis and Newman-Keul were done to compare age-related changes of cell type and adherence of reovirus. The variances were compared by Levy’s multiple comparisons.
Results Peyer’s patches were identified on the serosal surface of the antimesenteric side of the small intestine as a bulge or an area having a network of anastomosing vessels associated with lymphoid follicles. Peyer’s patches were easily seen with a dissecting microscope in mice 4 days of age and older but were only occasionally recognized in this manner in mice 2 days of age. In most 2-day-old mice, light microscopic examination of serial l-2-pm sec-
84
WOLF ET AL
tions was required to identify patches. From the 29 blocks of tissue that were studied from fifteen &dayold mice, 10 Peyer’s patches from 7 different mice were identified. The Peyer’s patches observed in 2-day-old mice, previously termed “anlage” patches (18), were small and appeared immature (Figure la). The most abundant cell types were large and medium-sized lymphocytes, although some macrophages were present. Very few small lymphocytes were seen. Between postnatal days 4 and 7, the size of Peyer’s patches and number of small lymphocytes in most Peyer’s patches gradually increased (Figure lb). All 10 immature Peyer’s patches identified in the &day-old mice were examined by transmission electron microscopy. Nine of 10 patches were from mice inoculated with reovirus. The majority (58%) of cells overlying domes of these immature patches were undifferentiated epithelial cells (Table 1). These cells had relatively short microvilli, a poorly developed terminal web, only sparse elements of the apical tubulovesicular system, many free ribosomes, and few elements of formed endoplasmic reticulum (Figure 2). In contrast, characteristic absorptive cells of these neonatal mice, which constituted 38.5% of the dome epithelium (Table l), had longer microvilli, a well-developed apical tubulovesicular system, fewer free ribosomes, and more formed endoplasmic reticulum (Figure 31. Tuft and goblet cells, although uncommon, were seen occasionally. A total of six M cells was identified in the epithelium overlying three of the 10 Peyer’s patches obtained from two 2-day-old mice. These M cells had structural features similar to those seen in older animals; both mature and immature forms were present (6-9,19). The M cells had short, irregular, branched and wide microvilli and a poorly developed terminal web. However, only one M-cell profile had the typical central hollow that contained mononuclear cells (6-9,19). In other M-cell profiles, mononuclear cells were adjacent to the basal cytoplasm, which they often indented (Figure 4). As the mice aged, the mean percentage of undifferentiated cells overlying Peyer’s patch domes decreased from 58% at 2 days of age to 41% at 7 days of age, whereas the percentage of absorptive cells increased from 38.5% at 2 days of age to 48% at 7 days of age (Table 1). The mean percentage of M cells increased from 0.9% of the cells overlying Peyer’s patch domes at 2 days to 7.4% at 7 days. The percentage of goblet and tuft cells remained essentially unchanged between 2 and 7 days after birth. However, in mice 2-6 days of age, there was a large variation from patch to patch in the percentage of undifferentiated cells in the epithelium overlying sectioned patch profiles. This variation occurred
GASTROENTEROLOGY
Figure
1
Vol. 92, No. 1
Low-magnification electron micrographs of lymphoid follicles of Peyer’s patches from a &day-old (a) and a 6-day-old (b) BALBicByJ suckling mouse. The dark lines represent the bars of the copper grid. The greater size and increased cellularity at 6 days is evident (X600).
REOVIRUS AND THE INTESTINE
January 1987
Table
1. Cellular
Composition
of the Epithelium
IN NEONATAL
85
MICE
of Peyer’s Patches Mouse age [days]
Undifferentiated cells” Absorptive cells” M cells” Goblet cells” Tuft cells” Total cells
2
4
57.8 ?z 8.3 38.5 -r-9.4 0.9 k 0.9 1.9 2 1.0 1.2 k 0.9 197
57.8 2 7.8 37.8 2 7.3 3.0 +- 1.5 1.5 2 0.8 0.2 2 0.2 301
6
5 48.8 2 39.0 ? 4.8 + 5.2 f 2.6 2 208
7.3 5.4 2.3 1.1 1.5
38.5 ? 51.5 k 3.6 k 5.9 + 0.5 + 254
7 7.8 8.0 1.6 1.7 0.5
41.2 k 47.8 t 7.4 t 2.1 t 1.7 t 205
1.2 2.8 2.7 1.1 0.6
a Mean percentage 2 SEM. The counted cells were from the following number of patches and (mice]: day 2,5 (4); day 4, 6 (4): day 5.4 (4); day 6, 6 (4);day 7, 6 (5).
even among patches from the same animal. This is not surprising, as the intestine does not mature at the same rate in all mice nor do all patches mature in synchrony in a given mouse. By 7 days, the variation in the percentage of undifferentiated cells overlying patches from different animals decreased from a mean SE of 7.8 in the 2-6-day-old animals to 1.2 in the 7-day-old animals. This difference in variance was statistically significant (p < 0.01). There was also substantial variation in the number and percentage of M cells in sectioned profiles of patches in young neonates. At 2 days of age, only one M cell (3.6% of dome cells on that patch) was detected in the five scored patches from 4 animals. In &day-old mice, the number of detectable M cells in scored patch profiles varied from 7.1% to 0%. At 7 days, >lO% of the cell profiles from patches of 2 mice were M cells, whereas in another mouse 22% and 2.5% of dome epithelial cell profiles in each of two patches were M cells. In 2 other ‘I-day-old mice, 0% and 1.9% of the dome epithelial cell profiles were M cells. As a result of this variation, differences in the percentage of undifferentiated, absorptive, and M cells in patch profiles of mice 2-7 days of age did not achieve significance when subjected to rigorous statistical analysis. Reovirus adhered to the apical surface of >50% of M-cell profiles in all mice except for 2, in which adherence to 0 of 1 M-cell profile in 1 mouse and to 40% of the M-cell profiles in the other was demonstrated. Of the six M-cell profiles visualized in 2-dayold patches, five had adherent reovirus on their apical surface. Reovirus adherence to undifferentiated cells overlying Peyer’s patches showed large variations from patch to patch and animal to animal. At 2 and 4 days, a mean of 53% k 13% and 56% +- 8% of undifferentiated cell profiles had adherent reovirus, whereas from 5 to 7 days the mean percentage with adherent reovirus was 31%-36.5% (Figure 5). These differences were not statistically significant. Reovirus adherence to absorptive cell profiles overlying Peyer’s patch domes varied from a mean of
15% at 2 days to 34% at 4 days, and then decreased to 11% at 7 days. Variation of adherence among
individual mice was large, resulting in large standard errors and considerable overlap between
Figure
2. Undifferentiated cell with adherent reovirus (arrows] on the apical surface located in the epithelium overlying the dome of a Peyer’s patch of a Z-day-old BALBicByJ mouse 30 min after intrajejunal inoculation with reovirus type 1. Note the numerous free ribosomes and the absence of elements of an apical tubulovesicular system (X 18,500).
86
WOLF ET AL.
Figure
3
Ileal epithelium from a Z-day-old BALB/cByJ mouse 1 h after intrajejunal inoculation with reovirus type 1. a. A villus absorptive cell with reovirus adherent to the microvilli (short arrows) and within apical vesicles (long arrows) (X10,800). b. Peyer’s patch absorptive cell cytoplasm with reovirus (arrow] within a lysosomelike body (X72,900).
GASTROENTEROLOGY
Vol. 92, No. 1
groups (Figure 5). In contrast to the adherence of reovirus to the majority of M-cell profiles, in only three patches from two 4-day-old animals and a single patch from one 5-day-old animal did >50% of absorptive cell profiles have adherent reovirus. Adherence of reovirus to the apical surfaces of the villus absorptive cell profiles showed large variations among young neonates with means of 36% + 20% at 2 days and 54% k 11% at 4 days of age. The number of villus absorptive cell profiles with adherent reovirus decreased dramatically by 7 days; only 3% + 2.8% had adherent reovirus. Because of the overall large variance, however, the decrease in adherence with age did not achieve significance when subjected to rigorous statistical analysis (Figure 5). The decrease in the variance of adherence at 7 days was significant (p < 0.01). Adherence of reovirus to the apical cell surface did not correlate with endocytosis of reovirions into the epithelial cytoplasm in 2-day-old mice. Evidence of endocytosis by undifferentiated cells of reovirus type 1 was not seen although virus adhered to sectioned profiles of the majority of these cells. In contrast, in 12 of the many hundreds of absorptive cell profiles surveyed, reovirus type 1 was found within vesicles and/or lysosomelike bodies in the apical cytoplasm. Ten of these overlaid villi and two overlaid Peyer’s patches (Figures 3a and 3b). No differences were detected in the structural features of those few absorptive cells that endocytosed type 1 reovirions and the many that did not. At 2 days of age, multiple sections of each of six M cells showed that two had endocytosed reovirus into the apical cytoplasm (Figure 4).Within the M-cell cytoplasm, reovirus was found in apical vesicles and in a multivesicular body. However, at 2 and 4 days of age 0.5and 1 h after inoculation, no reovirions were seen in the basal two-thirds of cytoplasm of any epithelial cell, in intercellular spaces, or in the underlying lamina propria. In mice older than 2 days, reovirus was found within the cytoplasm only of M cells except for a single absorptive cell profile from a &day-old mouse in which one virion was found in an apical vesicle close to the microvillus surface. It could not be determined whether this was an endocytic vesicle or whether it represented a tangentially sectioned pit that communicated with the intestinal lumen out of the plane of section. In mice 5 days of age and older, evidence that M cells both endocytosed and transported reovirus type 1 across the epithelium was observed. Reovirions were found on the M-cell apical surface, within M-cell cytoplasmic vesicles, and in the intercellular space underlying the M cell (Figure 6). No reoviruslike structures were seen in any of the
January
1987
REOVIRUS
AND THE INTESTINE
IN NEONATAL
MICE
87
FYigure 4. Ileal epithelium overlying the dome of an immature Peyer’s patch from a Z-day old BALB/cByJ mouse obtained 1 h after intrajejunal inoculation with reovirus type 1. Reovirus is adherent to the apical surfaces of a tuft (T) and M (M) cell, but not to the apical surface of an adjacent absorptive (A) cell. and is found within apical cytoplasmic vesicles of the M cell (arrows) (x11.800).
88
WOLF
ET AL.
GASTROENTEROLOGY
Vol. 92. No. 1
70
In .-2 > $ a
60 f 50
.
t
z ? g 40 a c .t 3 30 UY = S z
20
8
130 169 102 95
N= AGE (days I=
2
4
Undifferentiated
5
6
86
59
7
2
Cells on
4
5
Absorptive
Peyer ‘s Patches Figure
121 83
Peyer’s
133
97
6
7
Cells
on
Patches
277
-
293 285 35 I 240
24567 Absorptive
Cells on
Villi
5. Adherence of reovirus type 1 to undifferentiated cells and absorptive cells overlying the domes of Peyer’s patches absorptive cells overlying villi from 2-7-day-old BALBicByJ mice killed 0.5-l h after inoculation with reovirus type standard error of the mean is indicated. The total number (N) of each cell type counted at a given age is shown.
control mice that had phosphate-buffered stilled into the intestinal lumen.
saline
in-
Discussion We have previously documented that reovirus type 1 adheres selectively to the apical surface of the M cell and is transported through its cytoplasm in endocytic vesicles to the underlying Peyer’s patch lymphoid follicle from which it is then disseminated to distant organs in mice 10 days of age and older (3-5). However, it was not known whether this occurred in younger neonatal mice, the time at which reovirus type 1 grows most avidly in many tissues including the brain (11; Sharpe AH and Fields BN, unpublished data). As it also was not known when M cells first appear and first transport reovirus during intestinal morphogenesis in mice, we determined in suckling mice 2-7 days of age (a) whether M cells are present, (b) the pattern of reovirus type 1 adherence to intestinal epithelial cells, and (c) which epithelial cell types endocytose reovirus type 1.
and to 1. One
Small intestinal M cells are not present on villi but are confined to the epithelium overlying the dome of lymphoid follicles (6-9). In previous studies, rudimentary Peyer’s patches were observed in the small intestine of rats 2 days before birth (20,21) and at birth in several strains of mice including BALB/cJ mice (18,22-25). These patches in newborn mice are not fully developed; they contain endothelial, histiocytic, reticular, and large lymphoid cells but few small lymphocytes at birth. However, over the next 3-4 days, they become populated with large numbers of small lymphocytes (18,22-25). Compartmentalization of Peyer’s patches in suckling rats into well-defined follicular and interfollicular zones does not occur until 12-14 days after birth (26). Membranous epithelial cells have been noted overlying the bursa of Fabricius in 15day-old chicken embryos and at birth in rabbits (7). A single putative M cell has been illustrated overlying a submucosal lymphoid cell collection in the small intestine of a 17-wk-old human fetus (27), and M cells were reported to be present but were not illustrated overly-
lanuary 1987
Figure
REOVIRUS
AND THE INTESTINE
IN NEONATAL
6. Epithelium overlying a Peyer’s patch from a &day-old BALBkByJ mouse 1 h after inoculation with reovirus are seen adherent to the luminal surface of an M cell, within vesicles of the M cell, and within the extracellular (X 11,600). Inset: higher magnification of one of the extracellular virions (x93.400).
ing lymphoid follicles in the small intestine of newborn rats (20). Our findings confirm that aggregates of lymphoid cells that probably represent immature Peyer’s patches are present in the small intestine in the early neonatal period in mice. We also found that a small fraction (Cl%) of the epithelial cells that overlie these immature Peyer’s patches 2 days after birth are differentiated M cells but over the next 5 days of development, the mean frequency of M cells in dome epithelium increases over eightfold. This increase with age of the percentage of M cells on the apical surface of the domes of Peyer’s patches does not appear to occur in all animals of a given age at the same time. We noted considerable variation in Mcell frequency over patches among animals of the same age and even among individual patches from the same animal. Undifferentiated cells are the most abundant cell type overlying domes of immature patches in z-day-
MICE
89
type 1. Virions space (arrows)
old mice. However, as the patches mature over the next 5 days, the percentage of undifferentiated cells appears to decrease progressively as the percentage of absorptive cells increases. However, there is considerable variation among mice of the same age and among different patches from the same mouse, suggesting that maturation of the epithelium overlying patches does not proceed at the same rate and probably reflects the degree of maturity of each individual patch. Reovirus type 1 adheres selectively to M cells and a few tuft cells in lo-day-old suckling mice but not to absorptive or goblet cells overlying ileal Peyer’s patch domes (3,~). In adult mice, however, in addition to adherence of virions to most M cells and a few tuft cells, adherence of virions to a variable but small percentage of patch absorptive and goblet cells was observed (4). No adherence to villus epithelium was observed in either lo-day-old suckling or adult mice. In the present study, we found that reovirus
90 WOLFETAL.
type 1 adhered consistently to >50% of sectioned M-cell profiles in mice from 2 to 7 days old in all but two patches, one of which had only a single M cell. On the other hand, there was also adherence to a large percentage of sectioned profiles of undifferentiated cells overlying Peyer’s patches in young neonates. At 2 days, reovirus adhered to a small percentage of Peyer’s patch absorptive cells. At 4 days, adherence to absorptive cells was maximal both over patches and on villi. After 4 days, there was a tendency for the adherence of type 1 reovirus to absorptive cells to decrease with advancing age; however, this did not achieve statistical significance because there was substantial variation among animals of a given age. Again, differences in the rate of intestinal maturation from animal to animal and the presence of Peyer’s patches at various stages of development in animals of a given age may be responsible for these large variances. However, by 7 days, although virus adhered to the majority of sectioned M-cell profiles, mean adherence to sectioned absorptive cell profiles overlying patches and villi was only 10% and 3% respectively, strongly suggesting the emergence of increasing selectivity of type 1 reovirus adherence similar to that previously documented by us in lo-day-old suckling and adult mice (3,4,19). Several factors may influence the emergence of selective adherence of type 1 reovirus to M cells on dome epithelium in mice during the first 10 neonatal days. Undifferentiated cells, which bind reovirus type 1 in neonates and which retain some capacity to bind reovirus type 1 even in adult mice (Bye WA and Trier JS, unpublished data), ultimately disappear from the overlying dome epithelium as Peyer’s patches mature. Instead, they become segregated to the developing crypts that eventually encircle the domes of Peyer’s patches much as crypts encircle villi (19,28). That most patch and villus absorptive cells fail to bind reovirus type 1 by 7-10 days after birth may reflect alterations in the composition of the glycoconjugates of the apical glycocalyx. Changes in surface glycoconjugates may also account for the apparent difference in adherence to patch absorptive cells in 2- versus 4-day-old mice. It is well established that the glycocalyx of the brush border of intestinal absorptive cells in adult animals changes dramatically during normal maturation from precursor undifferentiated crypt cells (29). In suckling rats, the binding of lectins to glycoconjugates on the apical surface of villus absorptive cells changes dramatically during the first 2 wk of postnatal development, suggesting a change in the exposed sugar and sialic acid residues in the glycocalyx (30).Alterations in the incorporation of precursor sugars into glycoconjugates of the microvillus membrane of rat intestinal
GASTROENTEROLOGYVol. 92,No.l
epithelial cells have also been noted during postnatal development (31,32). Thus, although the glycoconjugates in the apical membrane of epithelial cells overlying the domes of Peyer’s patches in mice have not yet been directly examined during postnatal development, there is abundant evidence that epithelial cell surface glycoconjugates elsewhere in the small intestine of the closely related rodent, the rat, change dramatically with maturation during the early postnatal period. Factors other than glycocalyx composition may be important in determining reovirus adherence during intestinal development. Changes in the intraluminal environment during postnatal maturation of the stomach and small intestine may also influence the evolution of reovirus type 1 adherence to intestinal epithelial cells. Treatment of cultured L cells with pronase decreases binding of type 1 reovirus by -50% (33). Intraluminal trypsin activity is very low in the intestine of newborn rats but increases on day 4 (34). The gastric pH in rats decreases considerably as acid secretion increases during the first 10 postnatal days (35). Changes in protease activity and intraluminal pH may result in structural alterations of the brush border glycocalyx of intestinal epithelial cells or reovirus surface proteins altering viral adherence and infectivity (15,33,36,37). Endocytosis of reovirus type 1 by absorptive cells was observed with certainty only in 2-day-old mice, in which it was a rare phenomenon. However, whether endocytosis of reovirus type 1 into absorptive cells in 2-day-old mice reflects internalization of reovirions adherent to the absorptive cell apical membrane or nonspecific fluid phase endocytosis, which characterizes macromolecular uptake in mouse ileal absorptive cells from birth until weaning (131, was not established by our studies. Endocytosed virions were confined to vesicles and lysosomelike structures in the apical cytoplasm and were never observed in the basal two-thirds of absorptive cells, in intercellular spaces between absorptive cells, or in the underlying lamina propria. Whether these internalized reovirions would ultimately be degraded, replicated, or transported across the cell is unknown. However, from our morphologic studies, it seems unlikely that absorptive cells represent an important site for reovirus type 1 penetration of the ileal mucosa, especially in mice 24 days old. Although uptake of virions by absorptive cells more than 1 h after inoculation cannot be ruled out in these neonatal mice, our previous studies in mice ~10 days old (4)showed that reovirus type 1 was not endocytosed by absorptive cells within 4 h after inoculation. In contrast, M cells endocytosed reovirus type 1 at all ages examined and appear to be the major if not the only site for penetration of the intestinal mucosal barrier by this virus.
January
REOVIRUS
1987
The presence of M cells in Z-day-old mice suggests that very early in the postnatal period reovirus can be transported across the epithelial barrier of the intestine to the underlying lymphoid tissue. However, because the patch is immature, the ensuing immunologic response may differ in neonatal compared with older suckling and weaned mice and may permit the more exuberant growth of reovirus observed in tissues distant from the intestine during the early neonatal period (Sharpe AH and Fields BN, unpublished data). Hence, it appears likely that the differences in growth of reovirus type 1 observed in tissues distant from the intestine as the mouse ages reflect events occurring at a step beyond actual penetration of the epithelial barrier.
of virions. 17.
18.
19.
20.
21.
22.
References concept of poliomyelitis infection. Sci1. Bodian D. Emerging ence 1955;122:105-8. BR. Mims CA, Sambrook J, White DO. 2. Fenner F. McAuslan 1974: The biology of animal viruses. New York: Academic, 356-9. R, et al. Intestinal M cells: a 3. Wolf JL, Rubin DH, Finberg pathway for entry of reovirus into the host. Science 1981; 212:471-2. RS. Finberg R, Dambrauskas R, Fields BN, 4. Wolf JL, Kauffman Trier JS. Determinants of reovirus interaction with the intestinal M cells and absorptive cells of murine intestine. Gastroenterology 1983;85:291-300. RS, Wolf JL, Finberg R, Trier JS, Fields BN. The e1 5. Kauffman protein determines the extent of spread of reovirus from the gastrointestinal tract of mice. Virology 1983;124:403-10. 6. Wolf JL, Bye WA. The membranous epithelial (M) cell and the mucosal immune system. Ann Rev Med 1984;35:95-112. DE, Cooper MD. Pinocytosis by epithelium associ7. Bockman ated with lymphoid follicles in the bursa of Fabricius, appendix, and Peyer’s patches. An electron microscopic study. Am J Anat 1973;136:455-77. cell specialization within 8. Owen RL, Jones AL. Epithelial human Peyer’s patches: an ultrastructural study of intestinal lymphoid follicles. Gastroenterology 1974;66:189-203. 9. Owen RL. Sequential uptake of horseradish peroxidase by lymphoid follicle epithelium of Peyer’s patches in the normal unobstructed mouse intestine: an ultrastructural study. Gastroenterology 1977;72:440-51. 10. Margolis G, Kilham L. Hydrocephalus in hamsters, ferrets, rats and mice following inoculations with reovirus type 1: pathologic studies. Lab Invest 1969;21:189-98. 11. Tardieu M, Powers ML, Weiner HL. Age dependent susceptibility to reovirus type 3 encephalitis: role of viral and host factors. Ann Neural 1983;13:602-7. 12. Tardieu M, Weiner HL. Viral receptors on isolated murine and human ependymal cells. Science 1982:215:419-22. 13. Clark SL Jr. The ingestion of proteins and colloidal materials by columnar absorptive cells of the small intestine in suckling rats and mice. J Biophys Biochem 14. Sharpe AH, Fields BN. Reovirus
Cytol 1959;5:41-50. inhibition of cellular
RNA
and protein synthesis: role of the S, gene. Virology 1982; 122:381-91. 15. Drayna D, Fields BN. Genetic studies on the mechanism of chemical and physical inactivation of reovirus. J Gen Virol 1982;83:149-59. 16. Smith
RE, Zweerink
HJ, Joklik
WK. Polypeptide
components
AND THE INTESTINE
top
component
and
IN NEONATAL
cores
MICE
of reovirus
type
91
3.
Virology 1969;39:791-810. Karnovsky MJ. A formaldehyde-glutaraldehyde fixative of high osmolality for use in electron microscopy. J Cell Biol 1965;27:137A-8A3. Miller JJ, Studies of the phylogeny and ontogeny of the specialized lymphatic tissue venules. Lab Invest 1969:Zl: 484-90. Bye WA. Allan CH, Trier JS. Structure, distribution, and origin of M cells in Peyer’s patches of mouse ileum. Gastroenterology 1984;86:789-801. Wilders MM, Sminia T, Janse EM. Ontogeny of non-lymphoid and lymphoid cells in the rat gut with special reference to large mononuclear Ia-positive dendritic cells. Immunology 1983:50:303-31. Hummel KP. The structure and development of the lymphatic tissue in the intestine of the albino rat. Am J Anat 1935; 57:351-83. Waksman BH. The homing pattern of thymus-derived lymphocytes in calf and neonatal mouse Peyer’s patches. J Im-
munol 1973;111:878-84. SH. Weissman IL. Lymphoid tissue architecture: II. 23. Friedberg Ontogeny of peripheral T and B cells in mice: evidence against Peyer’s patches as the site of generation of B cells. J Immunol 1974;113:1477-92. 24. Joel DD. Hess MW. Cottier H. Thymic origin of lymphocytes in developing Peyer’s patches of newborn mice. Nature New Biol 1971:231:24-5. and pattern of 25. Joel DD, Hess MW. Cottier H. Magnitude thymic lymphocyte migration in neonatal mice. J Exp Med 1972;135:907-23. of Peyer’s patches 26. Sminia T, Janse EM, Plesch BEC. Ontogeny of the rat. Anat Ret 1983:207:309-16. cell types in the human fetal 27. Moxey PC, Trier JS. Specialized small intestine. Anat Ret 1978;191:269-86. in lymphoid 28. Bhalla DK, Owen RL. Cell renewal and migration follicles of Peyer’s patches and cecum. An autoradiographic study in mice. Gastroenterology 1982:82:232-42. morphology of the mucosa of 29. Trier JS. Madara JL. Functional the small intestine. In: Johnson LR. Christensen J, Grossman MI, Jacobson ED, Schultz SG. eds. Physiology of the gastrointestinal tract. New York: Raven, 1981:925-61. ML. Cell surface components of intes30. Etzler ME, Branstrator tinal epithelial cells and their relationship to cellular differentiation In: Development of mammalian absorptive processes. Ciba Foundation Symposium 70 (new series). New York: Excerpta Medica, 1979:51-68. A, Walker WA. Differences in microvil31. Bresson JL, Hercovics lus membrane glycoprotein fucosylation with age and differentiation (abstr). Gastroenterology 1982;82:1025. and transport of glycoproteins 32. Yen K-Y, Moog F. Biosynthesis in the small intestinal epithelium of rats: I. Developmental change and effect of hypophysectomy. Dev Biol 1984;lOl: 446-62. of 33. Epstein RL, Powers ML, Rogart RB, Weiner HL. Binding ‘*‘I-labeled reovirus to cell surface receptors. Virology 1984;133:46-55. P, Deschodt-Lanckman M. Camus J, Bruylands J. 34. Robberecht Christophe J. Pancreatic hydrolases from birth to weaning and dietary adaptation after weaning. Am J Physiol 1971; 221:376-81. HF. Ultrastructure and function of gastric parietal 35. Helander cells in the rat during development. Gastroenterology 1969;56:35-52. RS, Schaffer FL. Enzymatic enhancement of infec36. Spendlove tivity of reovirus. J Bacterial 1965;89:597-602. 37. Hauri H-P, Quaroni A, Isselbacher KJ. Biogenesis of intestinal plasma membrane: post-translational route and cleavage of sucrase-isomaltase. Proc Nat1 Acad Sci USA 1979;76:5183-6.